Ascendrc

refactor npu device manager (#31154)

* [NPU] Support executor with NPU

* Fix code according to reviews

* Fix code

* Add unittest for sub op npu

Fix reshape on GE graph

support parsing ascend rank table file

Destroy session first.

Add paddle ascend distribution training supported

Ascendrc add converted op : [range/equal/range/uniform_random/expand/squeeze], fix cast op bug  (#30797)

Ascendrc add converted op : [range/equal/range/uniform_random/expand/squeeze], fix cast op bug 

Merge ascend_optimizer and ascend_parser.

code style

[Feature] Build parser to support distributed training

cleanup test_ascend_group.py

Add startup bash files of test_ascend_group

Add Hccl program group

Pass device_ids info from launch to trainer

Build praser for Hcom* operators

Add distribution supported

* test=develop, fix fleet.metrics(mse, rmse, mae)

* add auc into 'all' list

* alias acc, expose to users

* update sample code

Co-authored-by: NseiriosPlus <tangwei12@baidu.com>

* test=develop, add distributed_infer

As the title

* skip quantizing ops in cpu inference, test=develop

As the title

Set expected place in child thread for dataloader to avoid costing cuda memory on other card (#30338)

* set expected place in child thread for dataloader

* set device id when set tensor from numpy

* revert tensor_py change

* add compile guard

* fix ci

* fix bug

* optimize memcpy perf for kunlun

* remove useless unitest for kunlun mean

* minor

* Implemented AddQuantDequantPass in imperative quantization.

* Supported LeakyReLU Quantization

* For meeting coverage rate.

* Changed the file name of test of AddQuantDequant

* Implemented more Quantized NoWeightLayers.

* Fix the loss cannot align problem between static and dynamic model quantization, add swish as supported quantized layer in imperative quantization.

* remove noweight_list

* support 2.0 API such as Pool2D and ReLu

fix compile problem on FT